Is it possible to make K-means clustering algorithm avoid some intensities

%% kmeans clustering
pixel_labels= imsegkmeans(maskedimage,2);
B = labeloverlay(maskedimage,pixel_labels);

This is code but it takes dark area of mask as group of pixels and segmenates it and makes rest of the image background, i only want it to segment image into 2 part without taking dark part into consideration (without zeros) I have to use k-value as 2, because some images don't have mask thank you

[Hari Krishna Ravuri]

Hi,

What is the meaning of the phrase "some images don't have mask"?

[Fatih Nurcin]

k-means clustering is applied to series of images, some images have mask and some not, this mask is applied before k-means in case of a specific color present in the image .

for k-means

Idea is to seperate image into 2 part without considering the dark pixels (which are masked due specific color present in the image, and not all the image have this color)

so i want k-means to ignore dark pixels, because if it doesn't ignore then k-means output will be same, it will not make any difference, segmentation will be mask (that was there before) and pixels

I want to ignore mask, and segment other pixels into two park as black and white

[Hari Krishna Ravuri]

Can you provide some sample normal images,sample images for which you want to apply mask before k-means and the images after applying the mask.